Optimizing content relay policy in publish-subscribe mobile social networks

Publish-subscribe mobile social networks enable content providers to disseminate up-to-date contents to end users with the help of mobile content relays by opportunistic wireless contacts. Since content providers, relays and end users are independent and self-interest entities in the mobile social networks, the content relay has to take a content requesting/transferring action to achieve the lowest cost. In this paper, we propose and solve a Markov decision process (MDP) based scheme for the content relay to optimally take the actions to receive contents from content providers, and to transfer/forward contents to end users. The relay takes an action based on the observed content price, the number of end users of contents, as well as the queue length. The proposed MDP scheme aims to minimize an expected cost of the content relay. The numerical results show that the proposed MDP scheme significantly outperforms baseline schemes.